Download Sulfate Complexation of Trivalent Lanthanides Proben by

Sulfate Complexation of Trivalent Lanthanides Proben
by NanoElectrospray Mass Spectrometry and TimeResolved Laser-Induced Luminescence
Thomas Vercouter,*,†,‡,§,$ Badia Amekraz,*,† Christophe Moulin,† Eric Giffaut,§ Pierre Vitorge†,‡
†
CEA-DEN Saclay DPC/SECR/LSRM, 91191 Gif-sur-Yvette Cedex, France, ‡ UMR 8587 (same
address), § Andra, 1/7 rue Jean Monnet, 92298 Châtenay-Malabry Cedex, France
* Authors to whom correspondence should be addressed. Phone: + 33-1-6908-7751. Fax: + 33-16908-5411. E-mail: [email protected]; [email protected].
RECEIVED DATE ()
1
ABSTRACT
Sulfate complexation of lanthanides is of great interest to predict the speciation of radionuclides in
natural environments. The formation of LaSO4+(aq) in HNO3/H2SO4 aqueous solutions of low ionic
strength (I) was studied by nanoElectrospray Ionization - Mass Spectrometry (nanoESI-MS). Several
gaseous species containing LaSO4+ were detected. The formation constant of LaSO4+(aq) was
determined and extrapolated to I = 0 (log β1° = 3.5±0.3) by using a simple Specific Ion interaction
Theory (SIT) formula. This value supports the potential of nanoESI-MS for the study of kinetically
labile species. The species La(SO4)2- was also detected. Besides, Time-Resolved Laser-Induced
Luminescence (TRLIL) was used to study Eu(III) speciation in the ionic conditions 0.02-0.05 M H+
(H2SO4/HClO4) and 0.4-2.0 M Na+ (Na2SO4/NaClO4). The data were interpreted with the species
EuSO4+ (log β1° = 3.78±0.1) and Eu(SO4)2– (log K2° = 1.5±0.2). For extrapolating to I = 0 all the
major ions were taken into account through several SIT ion-pair parameters, ε. Most of the ε values
were estimated by analogy to known parameters for similar ion-pair interactions using linear
correlations, while εEu3+,SO42- = 0.86±0.5 was fitted to experimental data, since, to date, SIT
coefficients between multi-charged species are not reported. The formation constants here obtained
confirm some of those previously measured for Ln(III) and An(III) by various experimental
techniques, and conversely do not give credit that in equilibrium conditions TRLIL and other
spectroscopic techniques would provide stability constants of only inner sphere complexes. The
fluorescence lifetimes measured for EuSO4+ and Eu(SO4)2– were consistent with the replacement of
one H2O molecule in the first coordination sphere of Eu3+ for each added SO42- ligand, suggesting a
monodentate SO42- coordination.
KEYWORDS : lanthanide, inorganic ligand, sulfate, complexation, electrospray, fluorescence.
2
1. Introduction
Owing to their extensive distribution in natural environments, and their binding capacity towards
metal ions, inorganic ligands play a major role in the environmental transport, fate and
bioavailability of heavy metals.1 This raises concern about the possibility of formation of soluble
complexes with inorganics, which could modify the migration of long-lived radionuclides released
in natural aquifers.2 The knowledge of radionuclide transport in the geosphere is a key issue for the
safety assessment of possible radioactive waste repositories.3 There is an interest in the
determination of thermodynamic data for their interactions with inorganics in order to properly
predict their speciation in natural systems.4 Several trivalent f-block elements represent a significant
part of the long-lived radionuclides, typically the actinides (An) Pu, Am, and Cm, and among the
lanthanides (Ln), the 151Sm isotope.5 Moreover analogies between An3+ and Ln3+ are sometimes used
to implement databases.6 Inorganic ligands can be roughly divided into two distinct groups based on
their reactivity for M3+ f-element cations and usual concentrations in groundwaters. The first
includes the carbonate and hydroxide anions that often form the major complexes with An3+ and
Ln3+ in deep groundwaters; they have been extensively studied.7-10 The second includes weaker or
less abundant ligands, sulfate, phosphate, silicate, chloride, and fluoride anions.11 Among these latter
ones, the sulfate anions deserve particular interest.11b In France, an underground laboratory for
radioactive waste disposal studies is currently under construction in a Callovo-Oxfordian clay
formation, where a sulfate concentration of 0.031 M has recently been proposed for the interstitial
waters of the clayey materials.12 Thus reliable complexation constants are needed to know whether
such a concentration could significantly affect the radionuclides speciation. In spite of many years of
research, sulfate complexation of An(III) and Ln(III) is still a matter of debate, and so on its
consequence on the mobilities of long-lived radionuclides through natural aquifers remains unclear.
Data for the formation of sulfate complexes of trivalent lanthanides have been obtained applying
techniques such as conductimetry, solvent extraction, and UV absorption. Table 1 summarizes the
stepwise formation constants (log β1 and log K2) reported in the literature for lanthanum13-23 and
3
europium13,22-31 trivalent ions. Several measurements in the ionic strength range 0.05-2 M have been
reported, while some values have been corrected to zero ionic strength and lie between 3.35 and 3.82
for log β1°, and 1.78 and 1.85 for log K2°. In the case of An(III) complexes, the formation constants
that are usually taken into account are the ones selected by the NEA-TDB critical reviews
(Thermochemical Data Base project of the Nuclear Energy Agency OECD):9,10 log β1° = 3.85±0.03
and log K2° = 1.5±0.7 have been selected by Silva et al. in 1995 from experimental data determined
by using solvent extraction, ion exchange and electromigration techniques;9 values for Cm(III) have
been provided by such techniques as well as by Time-Resolved Laser-Induced Luminescence
(TRLIL), and have been discussed in the recent NEA-TDB updated review.10 TRLIL has already
demonstrated its capacity to obtain reliable speciation data.32 Paviet et al. were the first to use
TRLIL in an attempt to directly observe the formation of sulfate complexes, and reported formation
constants for the complexes CmSO4+ and Cm(SO4)2– in 3 mol kg-1 NaCl/Na2SO4.33 In another study
by Neck et al., sulfate complexation of Cm(III) was investigated as a function of the ionic strength
(0-5.8 mol kg-1 NaCl/Na2SO4).34 However, the values derived from the spectroscopic studies on Cm
(log β1° = 3.30±0.15 and log K2° = 0.40±0.15) have been selected by the authors of the NEA-TDB
updated review, and appeared to be significantly lower than those previously selected for Am: 3.30
and 0.40 as compared to 3.85 and 1.5 for log β1° and log K2°, respectively.10 For selecting the new
Cm values, it has been argued that ion pairs had been misinterpreted as complexes in the previously
reported studies by solution-based methods, leading to incorrect larger formation constants.
However, we had already pointed out that spectroscopic techniques provide stability constants that
encompass the possible formation of ion pairs,8 and this will be reported here again; anyhow, in the
present study, the data for Eu(III) obtained by a spectroscopic technique (TRLIL) will be compared
with results from solution-based methods. The consistency of the analogy between Ln(III) and
An(III) in sulfate media will also be checked. It has also been argued in the NEA-TDB updated
review that the formation constants reported for Am(III) complexes with inorganics such as
carbonate, hydroxide, fluoride, and phosphate are close to, or smaller than those of the
corresponding U(VI) complexes.10,35 Thereby, the selected formation constants for U(VI) sulfate
4
complexes (log β1° = 3.15±0.02 and log K2° = 0.99±0.07) would discredit some values selected for
Am(III).9 However, the proposed correlation does not appear to hold for other ligands such as
chloride and nitrate, for which the selected formation constants for U(VI) complexes
(log β1°(UO2NO3+) = 0.30±0.15 and log β1°(UO2Cl+) = 0.17±0.02) are lower than the corresponding
Am(III) complexes (log β1°(AmNO32+) = 1.33±0.20 and log β1°(AmCl2+) = 0.24±0.03).10,35 These
observations likely indicate that making the hypothesis that a common trend for all ions would exist
is quite speculative, and therefore it cannot be taken as an indication of the reliability of the
spectroscopic data: Such comparisons and analogies are only rough guidelines. The effective charges
and ionic radii of Am3+ and U in UO22+ are indeed similar, but the coordination geometries are
different since the ligands bound to UO22+ are located in the plane perpendicular to the linear UO2axis.
In this study, thermodynamic methodologies were used together with the advanced spectrometric
techniques, TRLIL and nanoElectrospray Ionization - Mass Spectrometry (nanoESI-MS). To date,
the coordination chemistry of any lanthanide with sulfate has never been investigated by TRLIL.
The combination of spectroscopic information and measurements of species concentrations are much
valuable for a speciation purpose. A nanoESI - mass spectrometer was also used to observe the
formation of lanthanide sulfate complexes, and determine stability constants. This technique should
allow investigations of the aqueous speciation of many elements, providing rapid analyses without
constraining sample preparation.
In previous investigations, we have explored ESI-MS potential for metal speciation, and obtained
reliable formation constants for Ln(III) complexes with extractant molecules,36 as well as for
thorium hydroxides.37 However, there are currently two main difficulties encountered in the use of
ESI-MS as a means to assess stability constants. The first arises from the restricted ionic conditions
that can be investigated;38 the use of sodium salts to maintain a constant (and high) ionic strength in
the solutions was avoided, since it was observed to considerably alter the ESI-MS response. In fact,
almost all earlier metal complexation studies by ESI-MS were conducted using solutions of low
5
ionic strength, without any addition of electrolyte.39 The second difficulty comes from analysis of
the gas-phase species coming from Ln3+(aq), the “free” Ln(III) in aqueous solution. A few authors
provided some of the first examples of Ln(III) inorganic species observed by ESI-MS,40 indicating
that Ln(III) ions are strongly solvated in aqueous solutions (aquo ions Ln(H2O)n3+ (n = 8-9)), and
undergo gas-phase reduction in the transition from the condensed phase to the gas-phase leading to a
variety of ionic species, bare metal ions (Ln+ and/or Ln2+), oxides, hydroxides, or Ln3+ clusters. It is
noteworthy that, in aqueous solutions free of organic solvent, the ion intensity of Ln2+ becomes
important for the lanthanides higher than La, Ce and Pr.40b It follows a quite low total ion intensity,
due probably to a lower ion transmission efficiency of doubly-charged bare metal ion, when using a
quadrupole mass spectrometer. By contrast, the lower mass lanthanides La, Ce and Pr, which are
strong oxide formers, appear as oxide [LnO(H2O)n]+ or hydroxide [LnOH(H2O)n]2+ clusters in
spectra, and the total ion intensity could be analytically useful for the determination of the initial
aqueous concentrations.
In this study, we evaluated the use of nanoESI-MS as a means to obtain speciation information in
sulfate/Ln(III) aqueous solutions. The nanoESI process is based on a capillary action induced by an
applied electric field to draw the solution to the emitter tip.41 The solution flow rates are about 100
times lower than those used with ESI (generally 10 µl per minute with a syringe pump). In addition
to being more sensitive than conventional ESI, the spray is generated at lower temperature, voltage
and flow rate, which are favorable conditions where the gas-phase ions are representative of the
stability of the aqueous species. To our knowledge, this is the first report using a nanoESI-MS
approach to determine the formation constant of a metal complex. Herein we focused on the
monosulfate complex of La(III) formed in solutions at low ionic strength. Lanthanum was chosen
rather than higher mass lanthanides, to avoid the formation of Ln2+ bare metal ions from the aqueous
solutions and thereby to ensure analytically useful metal ion signals. Further investigations on
Ln(III) sulfate complexes were carried out by using TRLIL. Europium was chosen to take advantage
of its luminescence properties. Using TRLIL, the EuSO4+ and Eu(SO4)2– species were characterized
for their formation constants, as well as their first coordination sphere environment through lifetime
6
measurements. A speciation model is proposed and formation constants were determined from the
TRLIL data obtained in various ionic conditions. The Specific ion Interaction Theory (SIT)
formula9,42 was used for the description of the ionic medium/ionic strength dependence of the
activities (effective concentrations) of the species involved in the equilibrium reactions. The value of
the SIT coefficient εEu3+,SO42- was reported as the first experimental estimation to our knowledge for
a SIT coefficient between multi-charged species.
2. Experimental Section
Materials. Millipore deionized water (Alpha-Q, 18.2 MΩ cm) was used throughout the preparations.
The lanthanide solutions were prepared from La(NO3)3,6H2O (Prolabo, Rectapur®, 99.99%) and
Eu2O3 (Johnson Matthey, 99.99%). The perchloric, nitric and sulphuric acid concentrations were
adjusted by using 1 M stock solutions prepared from HClO4 70% (Merck, GR for analysis), HNO3
65% (Merck, Suprapur®) and H2SO4 98% (BDH, Aristar®), respectively, and all titrated with 0.1 M
NaOH (Merck, Titrisol®). NaClO4,H2O (R.P. Normapur™, >99.0%) and Na2SO4 (R.P. Normapur™,
>99.5%) were purchased from Merck and used without further purification.
Preparation procedures. All the preparations and measurements were performed at (23±1)°C.
NanoESI-MS measurements were performed in HNO3/H2SO4 solutions of La(III). Nitric acid was
used rather than perchloric acid which produced scattered MS signals due to isotopic effect of Cl. pH
was measured using a combined glass electrode (XC161, Radiometer Analytical) that was calibrated
for its linear response with commercial pH standards (Schott) with an estimated uncertainty of
±0.05. Since the ionic strength was low, typically 0.01-0.02 M, the effect of the junction potential
was neglected. H+ concentrations were deduced from the pH measurements that were corrected for
the activity coefficient of H+ calculated with the SIT formula (see Eq. 7). Two sets of experiments
have been done at pH ~2 with 10-3 and 5¥10-4 M La(NO3)3. Various volumes of a 0.01 M H2SO4
solution (pH = 1.83) were successively added to a 0.01 M HNO3 solution (pH = 2.02), both with the
same La(NO3)3 concentration, and pH was measured after each addition. Another set of experiments
7
was similarly performed by mixing 0.1 M HNO3 and 0.1 M H2SO4 solutions, both with 10-3 M
La(NO3)3. Since pH was out of the calibration range of the electrode, [H+] was calculated from the
initial concentration of the acids, the mass action law for the HSO4- dissociation, and the mass
conservation and electroneutrality relationships. [H+] was found to be close to 0.1 M when mixing
together the two solutions. In each set, the ratio of nitric acid to sulphuric acid was varied in order to
obtain increasing sulfate concentrations, while maintaining the ionic strength and pH roughly
constant, so that the compositions of the working solutions were: [SO42-] = 10-4-5.6¥10-3 M (1.83 <
pH < 2.02) with the ionic strength, I, varying from 0.01 to 0.02 M, and [SO42-] = 10-3-2¥10-2 M
(0.100 < [H+] < 0.092 M) with I from 0.10 to 0.14 M.
For TRLIL measurements, all the aqueous solutions were prepared with 10-4 M Eu(III) to keep
[Eu(III)] constant along the titrations. A first set of experiments was carried out at low ionic strength
by titrating a 0.01 M HClO4 solution with a 0.01 M H2SO4 solution, and pH was measured as for the
similar titration series in 0.01 M HNO3/H2SO4 solutions for nanoESI-MS measurements. Two other
sets of experiments were carried out at higher ionic strengths and pH > 3: Titrations were performed
using the initial Eu(III) solutions of 10-3 M HClO4 at I = 0.50 and 2.00 M (NaClO4), and 0.30 M
Na2SO4 as the titrant solution (-log[H+] = 3.9, I = 0.90 M). The H+ concentration of the Na2SO4
stock solution was determined by potentiometric measurements using an electrode whose reference
compartment was filled up with a 0.99 M NaClO4 / 0.01 M NaCl solution in order to minimize the
effect of the junction potential, and which was calibrated for H+ concentration with H+ buffer
solutions at I = 1 M. The ionic strength of the titrated solutions was calculated from the added
volumes and was ranging between 0.50 and 0.70 M, and 1.45 and 2.00 M for each series. As no
acido-basic reaction was expected during titration, [H+] was rather calculated and not measured in
order to limit systematic errors in potentiometric measurements due to the small variations of I.
Time-Resolved Laser-Induced Luminescence. Details about our “FLUO 2001” experimental setup have been given elsewhere.43 The main features of the excitation source are briefly given here as
it was different from that used in our previous studies. The excitation laser beam was generated by a
8
266 nm quadrupled Brilliant Nd-YAG laser, coupled to an optical parametric oscillator system
(Quantel, France). The wavelength was tuned to 395 nm, providing about 2 mJ of energy in a 5 ns
pulse with a repetition rate of 10 Hz. The data were treated using the OriginPro7 software
(OriginLab™).
NanoElectrospray Ionization - Mass Spectrometry. The mass spectra were recorded in positive
ion mode using a µ-Quattro triple-quadrupole spectrometer equipped with a nanoES interface
(Micromass, Manchester, UK). A 20-µL emitter tip was filled up with the solution to be analyzed,
and placed at 3 mm from the inlet orifice to the mass spectrometer (the optimal location that
maximizes the signal response); a voltage of 1.5 kV was supplied to the emitter tip to conduct
nanoelectrospray, providing a flow rate which has been determined to be about 0.1 µL/min. The
emitter tip was repositioned at its optimal location for each repeated analysis. The source
temperature was set to 80°C, and the sample cone voltage was set within the range 20-50 V. Spectra
were acquired at 6 s/scan over a mass range of m/z 50-1200 with an acquisition time of 3 min. For
MS/MS measurements, collision-induced dissociation of cluster ions was performed with argon; the
collision gas pressure was 2×10-3 mbar. Spectra were obtained at different collision energies ranging
from 5 to 30 eV.
3. Results and discussion
Thermodynamic equations. The stepwise formation constants for the monosulfate and disulfate
complexes of M3+ f-element are
+
β1 =
[MSO 4 ]
2−
[M 3+ ][SO 4 ]
(Eq. 1)
−
K2 =
[M(SO 4 ) 2 ]
+
2−
[MSO 4 ][SO 4 ]
(Eq. 2)
respectively. The stability of M(SO4)2- complex is equivalently defined by the overall stability
9
constant β2 = β1 K2. The formation constants extrapolated to zero ionic strength, K°, were calculated
with the SIT formula:
log K° = log Km – ∆z2D + Σi,j εi,j mj
(Eq. 3)
where the subscript m denotes the molality scale, and Km is related to K through molal-to-molar
conversion factors.42 D = 0.509Im1/2/(1 + 1.5Im1/2) is the Debye-Hückel term, Im is the ionic strength
(mol kg-1), ∆z2 is calculated from the charges of the species of the corresponding equilibrium: The
values are typically -12, -4 and -16 for β1, K2 and β2, respectively. εi,j is an empirical ion pair
interaction coefficient for the pair of species i and j; εi,j is assumed to equal zero for ions of same
charge-sign. Numerical values of εi,j were taken from the literature (Table 2) when available, or
obtained as explained below. mj is the molal concentration of the species j. The concentration of the
free SO42- ligand was calculated with Eq. 4 when the concentrations of Ln(III) sulfate complexes
were negligible in the mass balance of sulfate, that is for the Eu(III) experiments, where the effect of
metal complexation on [SO42-] was finally calculated to be less than 0.5%:
2−
[SO 4 ] =
[SO 4 ]0
1 + K b [H + ]
(Eq. 4)
where [SO4]0 is the total sulfate concentration.
-
Kb =
[HSO 4 ]
2[SO 4 ][H + ]
(Eq. 5)
is the basicity constant, and was calculated for each studied ionic medium using the SIT formula and
the appropriate εi,j coefficients (Table 2). For La(III) experiments, sulfate complexation was
accounted for in the mass balance leading to Eq. 9 (see below). The dependence of Kb with the ionic
strength had already been proposed by Grenthe et al. on the basis of four sets of experimental values
in solutions with NaClO4 as a supporting electrolyte;44 in this case Eq. 3 simplifies as,
log Kb,m = log Kb° - 4 D - ∆ε mNaClO4
(Eq. 6)
10
which was found to be reliable (Figure 1), and resulted in log Kb° = 1.989±0.084 and ∆ε =
(0.003±0.051) kg mol-1. In our work some of the solutions contained significant amounts of Na2SO4
salt, so that the simplified SIT formula (Eq. 6) was not well adapted. The SIT term was thus
developed to account for interactions with the main ions in the solutions, i.e. Na+, ClO4- and SO42for -log[H+] > 3. The corrected log Kb,m values slightly deviated from those in a pure NaClO4
medium when SO42- was not negligible against ClO4-, and for Im > 0.5 mol kg-1 (Figure 1).
Conversely, this correction was found to be negligible for the HClO4/H2SO4 and HNO3/H2SO4
working solutions for which Im < 0.1 mol kg-1 and -log[H+] > 1. Figure 1 also illustrates the case of
H2SO4 solutions without any supporting electrolyte: The calculated values deviate from the case
with NaClO4 only when Im > 0.3 mol kg-1, i.e. when -log[H+] < 0.5. In this case, the SIT term,
Σi,j εi,j mj (Eq. 3), had a much smaller effect on log Kb,m values than the Debye-Hückel contribution,
∆z2 D. For the experiments at pH lower than 2 and low ionic strength, a more significant effect
originated from the determination of the ligand concentrations from pH measurements according to
Eq. 4. Indeed, when pH was measured instead of -log[H+], Kb was corrected for γH+, the activity
coefficient of H+ calculated as
log γH+ = -D + εH+,ClO4- mClO4- + εH+,NO3- mNO3- + εH+,HSO4- mHSO4- + εH+,SO42- mSO42-
(Eq. 7)
We evaluated the unknown value of εH+,HSO4- by correlating available εM+,X- values with εNa+,Xpublished ones42 or calculated from the well-known corresponding Pitzer coefficients,45 we obtained
εH+,HSO4- = εNa+,HSO4- + (0.11±0.05). The H+ activity correction on log Kb ranged from 0.04 to 0.06,
hence from 0.02 to 0.03 on log[SO42-].
ESI-MS results for La(III). Ln(III) complexes are usually classified as kinetically labile on the
basis of fast formation and dissociation rate constants that reflect the strong ionic nature of f-element
bonding. The kinetics of LnSO4+ complexes has been examined by sound absorption techniques
from which stability constants were obtained for La and Eu (Table 1);18,25 these studies have shown
high rate constants for the formation of an inner sphere complex from the outer sphere complex (kf
11
1.0-3.4¥108 s-1), as well as for its dissociation (kd 2-7¥107 s-1). Nevertheless, there is increasing
acceptance that speciation of kinetically labile species can be maintained on the time scale of the ESI
process; for instance reliable formation constants were reported for metal complexes with kf ~105109 s-1.46 In previous studies, it has been shown that formation constants of lanthanide and actinide
complexes can be directly determined from quantitation of total metal speciation achieved by
monitoring gas-phase species coming from all free and complexed aqueous species.36,37 This
procedure was also used in the present study. The positive ion mode was required for the detection
of positively-charged ions induced by La3+(aq) and LaSO4+(aq). La(SO4)2-(aq) was however likely to
produce negatively-charged ions, for which a negative ion mode was required. Simultaneous
detection of positive and negative gas-phase ions is not possible; consequently it is also not possible
to make any direct comparison of the intensities measured in the two detection modes. For this
reason, experimental conditions were chosen so as to preferentially form La3+(aq) and LaSO4+(aq)
with negligible concentrations of the disulfate complex. Nevertheless, this latter species was formed
at higher sulfate concentrations obtained by increasing the H2SO4 concentration and thereby
decreasing the pH from about 2 to 1. The species identified in the nanoESI-MS spectra are
summarized in Table 3. MS/MS experiments involving collision-induced dissociation (CID) have
been used to probe the molecular ions generated by nanoESI (Table S1, Supporting Information).
The assignments reported in Table 3 agree with the detected daughter ions observed in representative
MS/MS spectra and the corresponding mass loss.
Quantitation of total La(III) speciation was achieved under relatively mild ion-source energy (ion
cluster mode). Figure 2 shows a representative nanoESI mass spectrum obtained in the positive ion
mode for an aqueous solution containing 10-3 M La(NO3)3 and a 2-fold molar ratio of SO42- at pH 2.
Whereas the [La(H2O)9]3+(aq) ion is known to be the predominant species in the aqueous phase at
pH 2, it was detected in spectra as oxides, hydroxides, and La3+ clusters (Table 3). The hydroxide
ions [LaOH(H2O)n]2+ and oxide ions [LaO(H2O)n]+ were observed. MS/MS spectrum confirmed the
assignments, where the loss of water molecules from the solvation shell is the major fragmentation
pathway for these ions (Table S1). The oxide ions are the predominant species throughout the
12
spectra. One can also observe that the high nitrate content of the solutions promotes the formation of
oxide ions that retain HNO3 during the gas-phase ion formation, such as [LaO(H2O)(HNO3)]+ and
[LaO(HNO3)]+. The likely species [La(NO3)2(H2O)n]+ that involve nitrate anions have also been
identified by MS/MS experiments. For instance, MS/MS spectra of the [La(NO3)2(H2O)]+ ion shows
that this ion readily lost H2O or decomposed under higher energetic conditions to give an oxide ion
[LaO(HNO3)]+. The analysis of these species to evaluate La(III) complexation by nitrate is not
straightforward. Because nitrate complexation is not very strong (log β1 slightly more than 1), the
association of the La3+ analyte ion with the not very volatile NO3- ligand may occur either in the
aqueous solution, or during the solution-to-gas phase transition leading to a non-specific binding
(cluster formation). The molar fraction of the LaNO32+(aq) species in the aqueous solution was
therefore calculated. Generally, f-elements are not expected to form aqueous complexes with
perchlorate anions, while inner sphere nitrato complexes have been proposed.47 Hence, complexation
by nitrate should be accounted for. This was determined from thermodynamic calculations using
data for the analog Am(III): log β1°(AmNO32+) = 1.33±0.20.10 The contribution of LaNO32+(aq) in
La(III) speciation was found to be less than 14% and 6% for pH 1 and 2, respectively. As indicated
below, this contribution has been taken into consideration in the analysis of the MS data. The weak
influence of nitrate complexes at pH 2 was also supported by the TRLIL results for equivalent
HNO3/H2SO4 solutions of Eu(III): The luminescence spectra did not significantly differ from those
for HClO4/H2SO4 solutions.
Concerning the monosulfate complex, LaSO4+(aq), the ions observed in the mass spectra
correspond to [La(SO4)(H2O)n]+, as well as mixed-solvent clusters, [La(SO4)(H2O)n(HNO3)]+ and
[La(SO4)(H2O)n(H2SO4)m]+. The effect of the non-specific binding of SO42- during the ES
desolvation process should be smaller than the complexation in the aqueous solution and was
thereby neglected. MS/MS spectra evidenced that these species dissociate by loosing solvent
molecules that surround the LaSO4+ ion (Table S1). Furthermore, it was found that under high
energetic conditions, fragmentation of the [La(SO4)(H2O)2]+ and [La(SO4)(H2O)2(HNO3)]+
complexes leads to the [LaO(H2O)]+ and [LaO(HNO3)]+ oxide species, respectively.
13
The efficiency of the conversion to ions in the gas-phase is likely to be similar for the various
species, as already observed from ion intensity measurements.38 The detection yields are also likely
to be similar for the monocharged gas-phase ions in the 120-500 m/z range. Consequently, the
summation of the ionic currents of the gas-phase species (Table 3) were assumed to be proportional
to the aqueous concentrations of either the free ion or the monosulfate complex. The ratio
R = [LaSO4+(aq)]/[La3+(aq)] was thereby calculated after [La3+(aq)] was corrected for the weak
nitrate complexation by subtracting the calculated concentration of LaNO32+(aq), which only slightly
increased the R values. The equilibrium constant β1 was determined from the mass action law
2−
log R = log β 1 + log[SO 4 ]
(Eq. 8)
and the equilibrium concentration of SO42-(aq)
 R 
[SO 4 ]0 − [La ]0 

2−
 R + 1
[SO 4 ] =
1 + K b [H + ]
(Eq. 9)
For 10-3 M and 5¥10-4 M La(III) solutions, plotting log R vs log[SO42-] gave a straight line of slope
+1 and intercept log β1 according to Eq. 8 (Figure 3). The slope +1 reflects the 1-1 stoichiometry of
the aqueous complex, which is good indication that we made reasonable assumptions for the
quantitative interpretation of MS data. For twice-diluted La(III) solutions at 5¥10-4 M, whereas two
of the dots were consistent with the model, two others significantly deviated, which was attributed to
very low ion intensities close to the detection limits, thus influencing the determination of R. In a
few solutions with the highest [SO42-], the formation of La(SO4)2-(aq) was suspected, and actually
detected by using the negative ion mode (Fig. S1, Table S2). Since the consumption of SO42- due to
the formation of La(SO4)2-(aq) was neglected in Eq. 9 because it could not be properly calculated,
the corresponding experimental dots were expected to deviate from the model towards elevated
[SO42-] as observed for [SO42-] > 0.01 M. This effect was even more stressed than expected, possibly
due to higher uncertainties on [La3+(aq)] that was determined from the peaks of low ion intensities;
the La3+(aq) concentration actually became lower than about 13% of the total lanthanum
14
concentration under these conditions.
Linear regression analysis of the nanoESI-MS data provided intercepts 3.0±0.2 and 2.9±0.3 for pH
1 and 2, respectively (±1.96¥σ, where σ is the standard deviation). Correction for nitrate
complexation was only significant for pH 1: log β1 = 3.1±0.3 (Figure 3). A good agreement between
the log β1 values was observed within uncertainties although a slight difference should be expected
due to the difference of ionic strengths. The log β1 values were extrapolated to I = 0 using the
simplified SIT formula
log β1° = log β1,m + 12 D + ∆ε Im
(Eq. 10)
as the influence of the ion pair term (∆ε Im) was small for I < 0.1 M. This term associated to the
complexation reaction was taken as: (εLaSO4+,HSO4- - εLa3+,HSO4-) mHSO4- - εH+,SO42- mH+ ≈ -0.06±0.17
(Table 2). This definition is consistent with a predominance of HSO4-, which is a rough
approximation because, beside HSO4-, the solutions also contained the NO3- and SO42- counteranions. However, the assumption was found to be relevant since the |∆ε Im| term was calculated to be
always less than 0.03, and did not significantly influence the calculations. Ionic strength corrections,
log β1° - log β1,m, were calculated from Eq. 10 and were found to equal 1.32±0.03 and 0.60±0.01 for
I = 0.1 and 0.01 M, respectively, which corresponds to pH 1 and 2, respectively. Hence, log β1° was
calculated as 4.4±0.3 and 3.5±0.3 from the two series of experiments. The too high log β1° value of
4.4±0.3 compared to other data (Table 1) was thought to result from non-specific binding during the
ESI desolvation process, since the sulfate concentrations in solutions at pH 1 were higher than those
in the solutions at pH 2. The value determined by nanoESI-MS from dilute solutions at pH 2 is
presented with other published values for La(III) in Table 1. Direct comparison is only possible for
the data extrapolated to I = 0 for which the agreement is good. The values obtained by different
workers from conductimetry are about 3.6 while a calorimetric study provided 3.5. Another value
reported in a potentiometric study, is about 0.3 log unit higher than the nanoESI-MS value. The
uncertainty of ±0.3 is larger than those previously proposed from other techniques (Table 1). This is
mainly due to the difficulty of making quantitative measurements for the free La(III) in aqueous
15
solutions without organic solvents. However, the quantitative agreement demonstrates the potential
of nanoESI-MS for kinetically labile species. Within its uncertainty, the MS value of 3.5±0.3 lies
between the data selected by Silva et al. (3.85±0.03) for Am(III)9 and the one (3.30±0.15) for
Cm(III),10 and cannot really help to discuss these values.
TRLIL results for Eu(III). The evolution of TRLIL spectrum is presented in Figure 4 for
increasing sulfate concentrations ([SO42-] = 0-0.2 M in Na2SO4/NaClO4 aqueous media at I = 0.500.70 M and 3 < –log[H+] < 3.9). Spectroscopic features of uncomplexed Eu(III) in aqueous
perchlorate medium have been emphasized in extensive studies of solution chemistry of
europium.32a,48 The TRLIL spectra obtained with solutions of Eu3+ in the presence of only
perchlorate anions present four characteristic bands centred at 593, 618, 650 and 700 nm
corresponding to radiative transitions from the 5D0 excited state to the 7F1, 7F2, 7F3 and 7F4 ground
state manifold, respectively. The strongest transitions are 5D0 7F1,2,4 while 5D0 7F3 is weaker
because it is forbidden according to Laporte’s selection rules. The 5D0 7F2 transition (electric
dipole) exhibits hypersensitivity and can be used as a luminescence probe for complexation
analyses; its intensity increases much more than those of other transitions upon complexation.
Interestingly, the non-degenerated 5D0 7F0 transition at 580 nm only occurs when the local
symmetry of Eu3+ is low, particularly when there is no inversion center, so it evidences inner sphere
complex formation.
While increasing the sulfate concentration, the hypersensitive transition peak at 618 nm changed
more significantly in intensity and position than other peaks (Figure 4); a slight shift (about 2 nm) of
its maximum towards the low wavelengths was observed. These spectral changes were attributed to
the formation of the sulfate complexes of Eu(III). The enhancements of the peaks at 593 and 700 nm
likely indicate that at least one of the Eu(III) species has either a higher luminescence quantum yield
or a higher absorption coefficient than Eu3+ at the 395 nm excitation wavelength. The detection of
the 5D0 7F0 emission at 580 nm is consistent with complex formation.
16
The quantitative analysis of TRLIL spectra was based on the intensity changes of the
hypersensitive transition peak. The measured intensity, Imes, was normalized (InormR) in relation to
[Eu]T, the total europium concentration, and I0°, the molar fluorescence intensity of Eu3+. As for
classical spectrophotometry, the change of the Eu(III) emission was described with the theoretical
expression:
I norm
R
=
I mes
=
[Eu ] T I 0 °
∑ (I i R β i [SO 4 2− ]i )
0 <i < 2
∑ (β i [SO 4
2− i
(Eq. 12)
] )
0< i < 2
where IiR = Ii° / I0° and Ii° is the molar fluorescence intensity of Eu(SO4)i3-2i. According to the SIT
formula, the dependences of the formation constants with the ionic strength are
log β1,m = log β1° - 12 D - ∆1(εm)
(Eq. 13)
log K2,m = log K2° - 4 D - ∆2(εm)
(Eq. 14)
The SIT terms, ∆1(εm) and ∆2(εm), are related to the interactions with the ionic components of the
solutions.49 Some of the ε values involved in ∆1(εm) and ∆2(εm) were available in the literature or
estimated by analogy to other M3+ cations, while εEu3+,HSO4-, εEu3+,SO42-, εEuSO4+,HSO4-, εEuSO4+,SO42- and
εH+,Eu(SO4)2- were unknown (Table 2). The determination of these latter parameters by curve fitting
technique turned out to be not relevant since their influence on ionic strength corrections was not so
high compared to a mean ∆ε value in a simplified SIT formula. Hence, we found it better to estimate
them using correlations as already proposed.44,50 For a given anion X-, εMz+,X- were found to correlate
linearly with z/rMz+ where rMz+ is the ionic radius of Mz+.51 Hence to determine εEu3+,HSO4-, we
calculated εK+,HSO4-, εMg2+,HSO4-, εCa2+,HSO4- and εFe2+,HSO4-, from Pitzer parameters for the
corresponding interactions.45 The linear regression applied to these four values and the tabulated
εNa+,HSO4-42 gave: εMz+,HSO4- = 0.186 (z/rMz+) – 0.196 (R2 = 0.93). We also obtained εMSO4+,SO42- =
0.205 (z/rMz+) – 0.331, but using only the two tabulated values for εLi+,SO42- and εNa+,SO42-.42 We
estimated the value of εEuSO4+,HSO4- as (εAmSO4+,ClO4- + εNa+,HSO4- - εNa+,ClO4-) = 0.20±0.10 from
17
tabulated data.42 This estimation is consistent with what would be found by Ciavatta’s method:52
εEuSO4+,HSO4- ≈ (εEu3+,HSO4- + εNa+,SO42-)/2 = 0.11±0.15. εH+,Eu(SO4)2- was not determined because of its
negligible effect under our conditions. Hence εEu3+,SO42- was the only fitted specific ion coefficient as
it involves multi-charged species for which correlations are not obvious, although it only had a weak
influence on the fit for [SO42-] > 0.1 M.
The luminescence spectra have been obtained for three series of titration experiments in different
ionic conditions: 0.02-0.05 M H+ (H2SO4/HClO4), 0.40-0.55 M Na+ and 2.00-1.30 M Na+
(Na2SO4/NaClO4). The intensities InormR at 618 nm are plotted against log[SO42-] in Figure 5. The
sensitivity of the analysis was assessed by examining several curve fits, where log β1 and log K2
were taken as functions of ionic media using Eqs. 13 and 14. The three more significant modelings
are represented in Figure 5(a). The aim was to determine the speciation model that best described the
data. When assuming the formation of EuSO4+ only, i.e. adjusting log β1° and I1R, the modeling
deviated from the data, except when log[SO42-] < -2.8 for the series with low ionic strength (Model
A). A better fit, very similar to Model B, was obtained when also adjusting εEu3+,SO42- that took the
value 6.9; this very high value is however unrealistic and only reflects the correlation between ε
parameters and stability constants. Model B was based on the assumption that EuSO4+ and Eu(SO4)2formed, and that both complexes had the same I°, which could be the case for instance if the second
SO42- do not enter into the first coordination sphere of Eu3+, but rather forms an outer sphere
complex, EuSO4+,SO42-. Details of the calculations are given in Appendix. The corresponding fitted
curves described fairly well the data; for the lowest ionic strength series, the few data dots at
log[SO42-] > -2.6 fell down the curve, whereas for the other two series, the curvature was too high to
perfectly match the data for -3 < log[SO42-] < -1.7. In Model C, log β1°, log K2°, I1R and I2R were
fitted, and the resulting curves better described the three sets of experiments. Model C was thus
found to be more relevant suggesting that EuSO4+ and Eu(SO4)2- formed, each species defined by a
specific IR value. The fit with Model C finally resulted in log β1° = 3.78±0.1 and log K2° = 1.5±0.2,
and the relative intensities at 618 nm, I1R = 2.6±0.1 and I2R = 5.6±0.3. A possible further complex
18
Eu(SO4)33- was insignificant under these conditions. The speciation diagrams are presented in Figure
5(b) for the three sets of ionic conditions. As expected, EuSO4+ was the major species (> 70%) at
low ionic strength (I = 0.02 M) in relation to the Eu3+ aquo ion that was better stabilized at higher I.
The formation of Eu(SO4)2- was observed when increasing the sulfate concentration and was even
the major complex (~60%) at medium ionic strength (I = 0.55 M).
The formation constants are reported in Table 1 with other published values for Eu(III). The values
at zero ionic strength show a good agreement between our TRLIL data and the ones obtained by
other techniques such as sound absorption, electrophoresis and solvent extraction, despite the
scattering of the log K2 values from the literature for a given ionic strength. Interestingly, our results
also agree well with the data for Am(III) and Cm(III) obtained by solution-based methods,9 but are
significantly higher than those for Cm(III) obtained by TRLIL.33,34 No explanation can be reasonably
offered for this latter observation, except that the studies were carried out in different ionic media
(NaClO4 vs NaCl). This however should not be responsible of such differences, unless medium
effects favored ion pairing, which is unlikely. Figure 6 illustrates the ionic strength dependence of β1
up to 2 M for a NaClO4 medium. The nanoESI-MS and TRLIL log β1 values as well as published
data for the La(III) and Eu(III) are consistent with the SIT formula for a NaClO4 electrolyte. Some of
our experimental values deviate because we accounted for short-range interactions with SO42- as in
the calculation of Kb. Thus, these data were naturally closer to the SIT curve corresponding to a pure
Na2SO4 medium. The fitted value of εEu3+,SO42- was found to be 0.86±0.5, and was associated with a
large uncertainty, since it only influenced the data for log[SO42-] > -1. Anyhow, we obtained the first
estimation to our knowledge for a SIT coefficient between multi-charged species.
The specific luminescence spectra for Eu3+, EuSO4+ and Eu(SO4)2- were determined by spectral
decomposition (Figure 7). The hypersensitive peak increases when the ligand binds Eu3+, with slight
differences of its shape in relation to different splitting effects of the 7F2 level. For EuSO4+, the peaks
at 593 and 700 nm are similar to that for Eu3+, suggesting that the yield of the radiative deexcitation
processes is the same, whereas the overall emission for Eu(SO4)2- is more intense, which is likely
due to better absorption at 395 nm. As already noted, the detection of fluorescence at 580 nm
19
reveals changes of the symmetry of the hydrated Eu(III) indicating the replacement of one or more
water molecules with one or two sulfate ions in the primary coordination sphere. The EuSO4+ and
Eu(SO4)2- species were characterized for their first coordination sphere environment through lifetime
measurements. As previously demonstrated, it is possible to correlate the primary hydration number
of europium (NH2O) and the lifetime of its 5D0 emitting level (τ).53 Such a correlation was reported by
Kimura and Choppin, NH2O = 1070/τ - 0.62, providing hydration numbers with an uncertainty of
±0.5.54 Indeed, the lifetime measured for Eu(III) in a 0.01 M HClO4 solution is 110±10 µs which
indicates the presence of nine water molecules in the internal coordination sphere of the Eu3+ aquo
ion, while a hydration number between 8 and 9 is expected.55 In all the solutions, the emission decay
was treated with a single exponential curve. The corresponding lifetimes slowly increased with the
formation of the mono and disulfate complexes. Hydration numbers were calculated from several
lifetime measurements (Table 4), and were interpreted as averages of the hydration numbers of Eu3+,
EuSO4+ and Eu(SO4)2- weighted by their concentrations in the solution. These hydration numbers
indicate the number of water molecules replaced by SO42- for each species. When the monosulfate
complex is predominant (0.01 M H2SO4 solution where metal speciation is: 19.4% Eu3+, 74.0%
EuSO4+, and 6.6% Eu(SO4)2-), τ was measured to be 123±10 µs, i.e. 8.1±0.5 remaining water
molecules. While increasing sulfate concentration up to 0.3 M, the predominant species is the
disulfate complex (0.3 M Na2SO4 solution where metal speciation is: 7.9% Eu3+, 36.8% EuSO4+, and
55.3% Eu(SO4)22-) and τ was 133±10 µs, i.e. 7.4±0.5 remaining water molecules. Considering
different possible hydration numbers for each species, the most reliable set of values was found to be
9, 8 and 7 for Eu3+, EuSO4+ and Eu(SO4)2-, respectively, according to the speciation results. This is
consistent with a mechanism where each sulfate molecule entering the internal coordination sphere
of europium is likely to exclude one water molecule from the primary hydration sphere, suggesting
that the sulfate ion acts as a monodentate ligand in aqueous solution. The same conclusion was
previously made from sound absorption measurements.25,56 The SO42- substitution rates with
lanthanides reported from these measurements are about 108 –107 s-1; since these values are close to
those found for water exchange, they have been interpreted as being indicative of the monodentate
20
nature of SO42- binding. For instance, the exchange rates of acetate (CH3COO-) substitution are two
orders of magnitude slower and were taken to be characteristic of a bidentate interaction.
The ratio of inner to outer sphere monosulfate complexes of lanthanides and actinides has been
previously examined.23,27 It is noteworthy that the interpretation of the luminescence spectra
obtained in this study does not exclude the formation of outer sphere complexes. Several authors
have discussed differences of stability constants of lanthanide complexes determined by
spectrophotometric and solution-based methods, and concluded that only the formation constant of
inner sphere complexes could be measured by spectrophotometric techniques.10,26 This belief was
found to be not consistent with thermodynamics, when equilibrium is achieved between inner and
outer sphere complexes.8 Despite spectroscopic changes are essentially due to inner sphere
complexes, the actually measured formation constant is the sum of the formation constants for the
inner and outer sphere species, β (tot) = β (in) + β (out), due to the equilibrium between the two
complexes (see Appendix). This conclusion was also confirmed by Hale and Spedding from their
UV absorption study dealing with the formation of EuSO4+.27 For instance, DeCarvalho and Choppin
have determined the formation constant of EuSO4+ in 2 M NaClO4 solutions by potentiometry
(log β1 = 1.37±0.08) and solvent extraction techniques (log β1 = 1.38±0.06);23 these values are in
good agreement with our TRLIL value at the same ionic strength (log β1 = 1.36±0.1).
Conclusion
Sulfate complexation of La(III) and Eu(III) has been investigated for the first time to our knowledge
by nanoElectrospray Ionization - Mass Spectrometry and Time-Resolved Laser-Induced
Luminescence. From the interpretation of luminescence lifetimes, the sulfate anion was concluded to
exchange with a single water molecule of the first coordination sphere, suggesting it is a
monodentate ligand towards trivalent f-element cations. NanoESI-MS provided a relevant stability
constant for the labile LaSO4+ complex, and confirmed its capacity to be a useful speciation tool for
studies of inorganic aqueous speciation of metal ions. However, we could only use this technique for
the characterization of species which were formed at low ionic strength. Stabilities of EuSO4+ and
21
Eu(SO4)2- were determined as functions of the ionic media by TRLIL. Our speciation model for
Eu(III) was consistent with earlier investigations by classical techniques and, interestingly, with
studies on Am(III) and Cm(III), suggesting a good analogy. This experimentally confirmed that
spectroscopic techniques do not provide stability constants for only inner sphere complexes, but
rather global constants for inner and outer sphere complexes, when existing. In equilibrium
conditions in interstitial waters of clayey materials of the Callovo-Oxfordian clay formation, the
ionic strength and the sulfate concentration had been estimated to be 0.1 M and 0.031 M,
respectively. Under these conditions, the concentration ratios of LnSO4+ and Ln(SO4)2- over Ln3+
were calculated to be 10.3 and 3.9, respectively, using the stability constants determined in this work
for Eu(III).
Acknowledgements
The authors are grateful to CEA DEN/DSOE (R&D) and ANDRA for financial support.
Appendix
The reaction of a ligand Ly- with an aquatic metal ion Mz+ is generally described by the EigenTamm mechanism,57 whereby the ultimate step is an equilibrium between inner (MiLjiz-jy
outer sphere (MiLjiz-jy
(in)
) and
(out)
) complexes as illustrated by Eq. A1 where δn denotes the hydration
number variation.
MiLjiz-jy (in) MiLjiz-jy (out) + δn H2O
(Eq. A1)
The thermodynamic constant that characterizes this equilibrium is:
ki,j= βi,j(out) /βi,j(in)
(Eq. A2)
where βi,j(in) and βi,j(out) are the formation constants of the inner and outer sphere complexes,
respectively. Thus, the measured intensity, Imes (light absorption or emission), writes
Imes = Σi,j (Ii,j°(in) [MiLjiz-jy (in)] + Ii,j°(out) [MiLjiz-jy (out)]
(Eq. A3)
where Ii,j°(in) and Ii,j°(out) are molar absorption coefficients or luminescence quantum yields. Eq. A3
22
also writes
Imes = Σi,j (βi,j(tot) Ii,j°(tot) [Mz+]i [Ly-]j)
(Eq. A4)
where we have noted
[M i L j
β i , j ( tot ) =
iz − jy ( in )
] + [M i L j
z+ i
iz − jy ( out )
y− j
[M ] [L ]
I i , j °( tot ) =
]
= β i, j
( in )
+ β i, j
( out )
β i , j(in ) I i, j °(in ) β i , j(out ) I i , j °(out ) Ii , j °(in ) I i, j °( out )
+
=
+
1 + ki, j 1 + 1
β i, j( tot )
β i , j( tot )
(Eq. A5)
(Eq. A6)
ki , j
which demonstrates that the formation constant measured for a species MiLjiz-jy is actually the sum of
the formation constants for the inner and outer sphere species. The MiLjiz-jy species is thus
characterized by Ii,j°(tot) and βi,j(tot), and not by Ii,j°(in) and βi,j(in). In this work, mono-nuclear species of
M3+ are involved (i = 1 and is omitted), and Eq. A4 results in an equation similar to that used for the
TRLIL data analysis (Eq. 12):
I mes
=
[M ] T
∑
0 < j< 2
(I j ° ( tot ) β j
∑
(β j
0 < j< 2
( tot )
( tot )
[ Ly − ] j )
[Ly − ] j )
(Eq. A7)
Even if spectral changes usually originate from the formation of inner sphere complexes, the
measured intensity probes the formation of both inner and outer sphere complexes, whose ratio is
actually constant, and equal to ki,j (Eq. A2). For instance, when only one ML3-y (out) complex is
formed, I1°(out) ≈ I0°, the specific intensity for the free M3+; then Imes ≈ I0° [M]T, so that
spectrophotometry is not sensitive to complexation at all. However, when ML3-y (in) is formed
exclusively or in addition to ML3-y (out), spectrophotometry is relevant for measuring either β1(tot) =
β1(in) or β1(tot) = β1(in) + β1(out), respectively.
Supporting Information Available: MS/MS data and assignments, MS data and assignments in the
negative ion mode. This material is available free of charge via the Internet at http://pubs.acs.org.
23
Footnotes
$
PhD grant from ANDRA.
(1) Bruno, J. Trace Element Modelling. In Modelling in Aquatic Chemistry; Grenthe, I.;
Puigdomenech, I. Eds; Elsevier Science BV: Amsterdam, 1997.
(2) Clark, D.L.; Hobart, D.E.; Neu, M.P. Chem. Rev. 1995, 95, 25-48.
(3) Hadermann, J. The pillars of Safety. In Modelling in Aquatic Chemistry; Grenthe, I.;
Puigdomenech, I. Eds; Elsevier Science BV: Amsterdam, 1997.
(4) Wanner, H. The NEA Thermochemical Data Base project, Report TDB-0.2, OECD Nuclear
Energy Agency, Data Bank, France, 1991.
(5) Johnson, L.H.; Shoesmith, D.W. Spent Fuel. In Radioactive waste forms for the future; Lutze,
W.; Ewing, R.C. Eds.; Elsevier Science Publishers B.V., 1988; Ch. 11, pp.635-698.
(6) (a) Choppin, G.R. J. Less-Common Metals 1983, 93, 232-330; (b) Krauskopf, K.B. Chem. Geol.
1986, 55, 323-335.
(7) Robouch, P. Contribution à la prévision du comportement de l'américium, du plutonium et du
neptunium dans la géosphère ; données géochimiques. Thesis, Université Louis Pasteur, Strasbourg,
France, 1987.
(8) Giffaut, E. Influence des ions chlorure sur la chimie des actinides. Effets de la radiolyse et de la
température. Thesis, Université Paris-sud, Orsay, France, 1994.
(9) Silva, R.J.; Bidoglio G.; Rand, M.H.; Robouch, P.B.; Wanner, H.; Puigdomenech, I. Chemical
Thermodynamics of Americium, Elsevier BV: Amsterdam, 1995.
(10) Guillaumont, R.; Fanghänel, T.; Fuger, J.; Grenthe, I.; Neck, V.; Palmer, D.A.; Rand, M.H.
Update on the Chemical Thermodynamics of Uranium, Neptunium, Plutonium, Americium and
Technetium, Elsevier BV: Amsterdam, 2003.
(11) (a) Wood, S.A. Chem. Geol. 1990, 82, 159-186; (b) Hendry, M.J.; Wassenaar, L.I. Water
Resour. Res. 2000, 36, 503-513.
(12) Gaucher, E.; Robelin, C.; Matray, J.M.; Négrel, G.; Gros, Y.; Heitz, J.F.; Vinsot, A.; Rebours,
H.; Cassagnabère, A.; Bouchet A. Water Geochemistry and Hydrogeology 2004, 29(1), 55-77.
(13) Izatt, R.M.; Eatough, D.; Christensen, J.J.; Bartholomew, C.H. J. Chem. Soc. A 1969, 45-47.
(14) Jenkins, I.L.; Monk, C.B. J. Am. Chem. Soc. 1950, 72, 2695-2698.
(15) Spedding, F.H.; Jaffe, S. J. Am. Chem. Soc. 1954, 76, 882-884.
(16) Fisher, F.H.; Davis, D.F. J. Phys. Chem. 1967, 71(4), 819-822.
24
(17) Farrow, M.M.; Purdie, N. J. Solution Chem. 1973, 2(6), 503-511.
(18) Reidler, J.; Silber, H.B. J. Phys. Chem. 1973, 77(10), 1275-1280.
(19) Jones, H.W.; Monk, C.B. Trans. Faraday Soc. 1952, 48, 929-933.
(20) Aziz, A.; Lyle, S.J. J. Inorg. Nucl. Chem. 1970, 32, 1925-1932.
(21) Simpson, C.; Matijevic, E. J. Solution Chem. 1987, 16(5), 411-417.
(22) Sekine, T. Acta Chem. Scand. 1965, 19, 1469-1475.
(23) DeCarvalho, R.G.; Choppin, G.G. J. Inorg. Nucl. Chem. 1967, 29,725-735.
(24) Laurie, S.H.; Monk, B. J. Chem. Soc. 1963, 3343-3347.
(25) Farrow, M.M.; Purdie, N. J. Solution Chem. 1973, 2(6), 513-523.
(26) Barnes, J.C. J. Chem. Soc. 1964, 3880-3885.
(27) Hale, C.F.; Spedding, F.H. J. Phys. Chem. 1972, 76(20), 2925-2929.
(28) Manning, P.G.; Monk, C.B. Trans. Faraday Soc. 1962, 58, 938-941.
(29) Aziz, A.; Lyle, S.J.; Naqvi, S.J. J. Inorg. Nucl. Chem. 1968, 30, 1013-1018.
(30) Bilal, B.A.; Koss, V. J. Inorg. Nucl. Chem. 1980, 42, 1064-1065.
(31) Bansal, B.M.L.; Patil, S.K., Sharma, H.D. J. Inorg. Nucl. Chem. 1964, 26, 993-1000.
(32) (a) Bünzli, J.C.G. Luminescent probes. In Lanthanides Probes in life, Chemical and Earth
Science – Theory and Practice, Bünzli, J.C.G., Choppin, G.R., Eds; Elsevier: Amsterdam, 1989; (b)
Moulin, C.; Decambox, P.; Moulin, V.; Decaillon, J.C. Anal. Chem. 1995, 67, 348-353; (c) Colette,
S.; Amekraz, B.; Madic, C.; Berthon, L.; Cote, G.; Moulin, C. Inorg. Chem. 2004, 43, 6745-6751.
(33) Paviet, P.; Fanghänel, T.; Klenze, R.; Kim, J. I. Radiochim. Acta. 1996, 74, 99-103.
(34) Neck, V.; Fanghänel, T.; Kim, J. I. Aquatic Chemistry and thermodynamic modeling of trivalent
actinides, Wiss. Ber. – Forschungszent. Karlsruhe, 1-108, 1998.
(35) Grenthe, I.; Fuger, J.; Konings, R.J.M.; Lemire, R.; Muller, A.B.; Nguyen-Trong, C.; Wanner,
H. Chemical Thermodynamics of Uranium, Elsevier BV: Amsterdam, 1992.
(36) (a) Colette, S.; Amekraz, B.; Madic, C.; Berthon, L.; Cote, G.; Moulin, C. Inorg. Chem. 2002,
41, 7031-7041; (b) Colette, S.; Amekraz, B.; Madic, C.; Berthon, L.; Cote, G.; Moulin, C. Inorg.
Chem. 2003, 42, 2215-2226.
(37) Moulin, C.; Amekraz, B.; Hubert, S.; Moulin, V. Anal. Chim. Acta 2001, 441, 269-279.
(38) Kebarle, P. J. Mass Spectrom. 2000, 35, 804-817.
25
(39) (a) Renaud, F.; Piguet, C.; Bernardinelli, G.; Hopfgartner, G.; Bünzli, J.-C. G. Chem. Commun.
1999, 457-458; (b) Ross, A.R.S.; Ikonomou, M.G.; Orians, K.J. Anal. Chim. Acta 2000, 411, 91-102;
(c) Delangle, P.; Husson, C.; Lebrun, C.; Pécaut, J.; Vottéro P.J.A. Inorg. Chem. 2001, 40, 29532962; (d) Muller, G.; Bünzli, J.-C. G.; Schenk, K. J.; Piguet, C.; Hopfgartner, G. Inorg. Chem. 2001,
40, 2642-2651.
(40) (a) Blades, A.T.; Jayaweera, P.; Ikonomou, M.G.; Kebarle, P. Int. J. Mass Spectrom. Ion
Processes 1990, 101, 325-336; (b) Stewart, I.I.; Horlick, G. Anal. Chem. 1994, 66, 3983-3993.
(41) (a) Wilm, M.S.; Mann, M. Int. J. Mass Spectrom. Ion Processes 1994, 136, 167-180; (b) Wilm,
M.S.; Mann, M. Anal. Chem. 1996, 68, 1-8; (c) Wood, T.D.; Moy, M.A.; Dolan, A.R.; Bigwarfe,
P.M.; White, T.P.; Smith, D.R.; Higbee, D.J. Appl. Spectrosc. Rev. 2003, 38, 187-244.
(42) Lemire, R.; Fuger, J.; Nitsche, H.; Potter, P.; Rand, M.; Rydberg, J.; Spahiu, K.; Sullivan, J.;
Ullman, W.; Vitorge, P.; Wanner, H. Chemical Thermodynamics of Neptunium and Plutonium,
Elsevier BV: Amsterdam, 2001.
(43) Plancque, G.; Moulin, V.; Toulhoat, P.; Moulin, C. Anal. Chim. Acta, 2003, 478, 11-22.
(44) Grenthe, I.; Plyasunov, A. V.; Spahiu, K. Estimations of Medium Effects on Thermodynamic
Data. In Modelling in Aquatic Chemistry; Grenthe, I.; Puigdomenech, I. Eds; Elsevier Science BV:
Amsterdam, 1997.
(45) Pitzer, K.S. Activity coefficients in electrolyte solutions, Pitzer, K.S., Ed.; 2nd edn., CRC Press,
1991.
(46) Wang, H.; Agnes, G.R. Anal. Chem. 1999, 71, 4166-4172.
(47) Sadowski, P.; Majdan, M. Monatsh. Chem. 1995, 126, 863-870.
(48) Carnall, W.T. The absorption and fluorescence spectra of rare earth ions in solution. In
Handbook on the Physics and Chemistry of Rare Earth, North Holland Publishing: Amsterdam,
1976.
(49)
∆1(εm)
=
(εEuSO4+,ClO4--εEu3+,ClO4-)mClO4-
+
(εEuSO4+,HSO4--εEu3+,HSO4-)mHSO4-
+
(εEuSO4+,SO42--εEu3+,SO42-)mSO42- - εNa+,SO42- mNa+ - εH+,SO42- mH+
∆2(εm) = (εNa+,Eu(SO4)2--εNa+,SO42-)mNa+ + (εH+,Eu(SO4)2--εH+,SO42-)mH+ - εEuSO4+,ClO4- mClO4- - εEuSO4+,HSO4mHSO4- - εEuSO4+,SO42- mSO42(50) Offerlé, S.; Capdevila, H.; Vitorge, P. Report CEA-N-2785, 1995.
(51) Choppin, G.R.; Rizkalla, E.N. Solution chemistry of actinides and lanthanides. In Handbook on
the Physics and Chemistry of Rare Earths. Lanthanides/Actinides: Chemistry, Gschneidner, K.A. Jr.;
26
Eyring, L.; Choppin, G.R.; Lander, G.H. Eds.; North-Holland: Amsterdam, 1994; Vol. 18, pp.559590.
(52) Ciavatta, L. Ann. Chim. (Rome) 1990, 80, 255-263.
(53) Horrocks, W.DeW.; Sudnick, D.R. J. Am. Chem. Soc. 1979, 101:2, 334.
(54) Kimura, T.; Choppin, G.R.; Kato, Y.; Yoshida, Z. Radiochim. Acta 1996, 72, 61.
(55) Rizkalla, E.N.; Choppin, G.R. Lanthanides and actinides hydration and hydrolysis. In
Handbook on the Physics and Chemistry of Rare Earths. Lanthanides/Actinides: Chemistry,
Gschneidner, K.A. Jr.; Eyring, L.; Choppin, G.R.; Lander, G.H. Eds.; North-Holland: Amsterdam,
1994; Vol. 18, pp.529-558.
(56) Margerum, D.W.; Cayley, G.R.; Weatherburn, D.C.; Pagenkopf, G.K. ACS Monograph 1978,
174 (Coord. Chem., v2), 1-220.
(57) Eigen, M.; Tamm, K. Z. Electrochem. 1969, 66, 93, 107.
27
Table 1. Stepwise formation constants of LaSO4+, La(SO4)2-, EuSO4+ and Eu(SO4)2- at 25°C.
Method(a)
Medium
log β1
I/M
3+
La +
cal
con
con
con
con
ul abs
pot
ESI-MS
extr
cal
extr
pot
HClO4/(Me4N)2SO4
La2(SO4)3
La2(SO4)3
La2(SO4)3
La2(SO4)3
H2SO4
HNO3/H2SO4
NaClO4
HClO4/NaClO4
NaClO4
NaClO4
0
0
0
0
0
0
0
I0
0.5
1
1
2
SO42
LaSO4
3.50±0.04
3.62
3.62
3.62
3.65±0.02
3.62
3.82±0.04
3.5±0.3
1.77±0.02
0.8
1.45±0.07
1.29±0.04
-
log K2
+
+
2--
Ref.
LaSO4 + SO4 La(SO4)2
1.85±0.07
-
0.89±0.01
0.2
1.01±0.08
-
13
14
15
16
17
18
19
this work
20
21
22
23
-
Eu3+ + SO42 EuSO4+
EuSO4+ + SO42 Eu(SO4)2
sol
0
3.72
24
ul abs
Eu2(SO4)3
0
3.66
25
0
3.54±0.03
1.78±0.09
13
cal
HClO4/(Me4N)2SO4
sp
NaClO4
0
3.35
26
sp
0
3.67±0.01
27
sp
NaClO4
0.046
2.76±0.01
1.26±0.25
27
NaClO4
extr
0
3.56
28
NaClO
extr
0.05
2.53
28
4
NaClO4
0.1
2.23
28
extr
extr
NaClO4
0.5
1.88±0.01
0.91±0.02
29
ix
NaClO4
0.5
1.87±0.01
0.86±0.02
29
extr
NaClO4
1
1.54±0.06
1.15±0.06
22
extr
NaCl
1
1.53±0.04(b)
30
ix
NaClO4
1
1.57±0.03
0.83±0.06
31
ix
HClO4
1
1.23±0.03
0.47±0.10
31
pot
NaClO4
2
1.37±0.08
0.59±0.10
23
extr
NaClO4
2
1.38±0.06
0.60±0.12
23
TRLIL
I0
3.78±0.06
1.5±0.2
this work
this work
H2SO4/HClO4
0.02-0.04
2.90-3.06
1.21-1.26
this work
Na2SO4/NaClO4 0.50-0.59
1.71-1.76
0.82-0.88
this work
Na2SO4/NaClO4 0.60-0.70
1.67-1.72
0.70-0.82
this work
Na2SO4/NaClO4
0.91
1.62
0.62
this work
Na2SO4/NaClO4 1.91-2.10
1.35-1.37
0.86-0.91
this work
Na2SO4/NaClO4 1.51-1.62
1.41-1.43
0.75-0.78
(a) cal = calorimetry, con = conductimetry, ul abs = ultrasonic absorption, pot = potentiometry, extr = solvent extraction,
sol = solubility, sp = spectrophotometry, ix = ion exchange.
(b) As the speciation model accounted for EuSO4+ and Eu(SO4)33-, the log β1 value may be influenced by log β3.
28
Table 2. SIT coefficients at 25°C.
εH+,ClO4-
Value(a)
/ kg mol-1
0.14±0.02
εH+,NO3-
0.07±0.01
Method
Ref.
42
εH+,HSO4- 0.10±0.06
εNa+,HSO4- + (0.11±0.05)
εH+,SO42- -0.03±0.06
≈ εLi+,SO42-
42
this
work
42
εNa+,HSO42- -0.01±0.02
42
εNa+,SO42- -0.12±0.06
42
εLa3+,ClO4- 0.47±0.03
εEu3+,ClO4- 0.49±0.03
εLa3+,HSO4- 0.28±0.14
0.47 < εLn3+,ClO4- < 0.52
42
0.186 (z/rMz+) - 0.196 (b)
this
work
εEu3+,HSO4- 0.33±0.14
this
work
0.22±0.09
42
≈ εAmSO4+,ClO4this
0.20±0.10 εAmSO4+,ClO4- - (0.02±0.02)
work
εEu3+,SO42- 0.86±0.5
εMSO4+,ClO4εMSO4+,HSO4-
from TRLIL data
εLaSO4+,SO42- -0.15±0.23
0.205 (z/rM3+) - 0.331
εEuSO4+,SO42- -0.14±0.25
this
work
42
εNa+,M(SO4)2- -0.05±0.07
≈ εNa+,Am(SO4)2(a) Uncertainty depends on the estimation method: For analogy, (σ2+0.052)0.5 kg mol-1 where σ is the original
uncertainty; for correlation, it is calculated from the standard error of parameters in the linear regression.
(b) Correlation on the basis of ε values calculated from Pitzer parameters:45 εK+,HSO4- = -0.04±0.04, εMg2+,HSO4- =
0.33±0.05, εCa2+,HSO4- = 0.12±0.05, εFe2+,HSO4- = 0.38±0.11.
29
Table 3. Complexes detected by nanoESI-MS for 10-3 M and 5×10-4 M La(NO3)3 under pH range 1-
2 in HNO3/H2SO4 medium.
La3+
LaSO4+
[LaOH(H2O)n]2+, n=5-7
m/z 123, 132, 141
[La(SO4)(H2O)n]+, n=1-4
m/z 253, 271, 289, 307
[LaO(H2O)n]+, n=0-2
m/z 155, 173, 191
[La(SO4)(H2O)n(HNO3)]+, n=0-3
m/z 298, 316, 334, 352
[LaO(H2O) n(HNO3)]+, n=0-1
m/z 218, 236
[La(SO4)(H2O)n(H2SO4)]+, n=0-4
m/z 333, 351, 369, 387, 405
[La(NO3)2(H2O)n]+, n=0-2
m/z 263, 281, 299
[La(SO4)(H2O)n(H2SO4)2]+, n=1-3
m/z 449, 467, 485
Table 4. Speciation results and measured fluorescence lifetimes of Eu(III) aqueous solutions at
25°C.
[Na+] / M log Kb
0.00
0.50
1.94
0.50
1.89
0.50
1.84
0.50
0.51
0.51
0.52
0.00
1.41
1.30
0.53
0.40
0.55
0.50
0.60
1.30
1.15
1.30
1.14
1.29
1.14
1.29
1.29
1.28
1.26
1.65
1.12
1.11
1.23
1.26
1.20
1.18
1.11
-log[H+] log[SO42-]
2.00
3.07
3.07
3.09
3.10
3.10
3.13
3.12
3.13
3.17
3.24
1.83
3.40
3.46
3.36
3.35
3.47
3.62
3.89
-6.71
-1.98
-2.12
-1.68
-1.94
-1.49
-1.73
-1.66
-1.46
-1.20
-2.22
-0.90
-0.83
-0.98
-1.00
-0.84
-0.70
-0.52
% of species
3+
Eu
100.0
100.0
79.0
67.4
63.9
58.0
52.1
45.5
41.2
30.3
19.4
19.4
19.3
16.6
13.2
10.9
10.5
8.1
7.9
EuSO4+
0.0
0.0
19.0
30.5
30.0
38.1
36.6
46.8
49.3
53.8
53.7
74.0
42.4
41.5
49.4
50.5
45.4
41.4
36.8
Eu(SO4)20.0
0.0
2.0
2.1
6.2
3.9
11.2
7.6
9.5
15.9
26.8
6.6
38.3
41.9
37.4
38.6
44.0
50.4
55.3
τ / µs
ΝΗ2Ο
110
113
112
117
113
118
119
120
121
123
126
123
127
129
129
130
132
133
133
9.0
8.8
8.9
8.5
8.8
8.4
8.4
8.3
8.2
8.1
7.9
8.1
7.8
7.7
7.7
7.6
7.5
7.4
7.4
30
Captions for figures
Figure 1. Dependence of log Kb,m with ionic strength, Im at 25°C. The thin and bold continuous
lines represent fits to experimental values (open circles) selected in Refs. 44 and 42, respectively,
with a simplified SIT formula for a NaClO4 medium (Eq. 6). Values were also calculated with the
SIT formula for each of our solutions accounting for the proportions of ionic constituents (cross and
black circles). For comparison, extrapolation of log Kb,m to high ionic strengths is shown (dotted
line) for a H2SO4 medium where HSO4- predominates.
Figure 2. NanoESI-MS spectra of 10-3 M La(NO3)3 and a 2-fold molar ratio of SO42- at pH 2,
HNO3/H2SO4 medium, cone-voltage 30 V.
Figure 3. Interpretation of nanoESI-MS results with the formation of LaSO4+. Experimental values
of log R and log β1 are represented against log[SO42-], calculated using Eqs. 8-9. Thin and bold
straight lines result from linear regression analyses of the experimental data for pH 1, 0.10 < I < 0.14
(■), and pH 2, 0.01 < I < 0.02 (▲) respectively. Close and open symbols refer to 10-3 M and
5¥10-4 M La(III) solutions, respectively.
Figure 4.
TRLIL spectra of Eu(III) with -4.1 < log[SO42-] < -1.6, in Na2SO4/NaClO4 aqueous
solutions with I = 0.5 M at –log[H+] > 3 and 25°C.
Figure 5. TRLIL data analysis at 25°C: (a) normalized relative intensity, InormR, at 618 nm against
log[SO42-], measured for Eu(III) aqueous solutions with different ionic conditions; the theoretical
curves are fitted to the data according to three different models (see text): Assuming the formation of
EuSO4+ from Eu3+ (model A), and adding Eu(SO4)2- as outer (model B) or outer and inner (model C)
complexes; (b) speciation diagrams of Eu(III) for the 3 sets of ionic conditions.
Figure 6. Experimental values of log β1 against ionic strength, Im, at 25°C for the formation of
LaSO4+ and EuSO4+. The data determined in this work (closed symbols) are compared to literature
data reported in Table 3 (open symbols). The dependence with Im is calculated using the simple SIT
formula (Eq. 10) and εi,j values (Table 2) when NaClO4 (continuous line) or Na2SO4 (dotted line)
predominates as supporting electrolytes.
Figure 7. TRLIL spectra of Eu3+, EuSO4+ and Eu(SO4)2- for λexcitation = 395 nm at 25°C.
31
Figures
Figure 1.
2
H+ + SO4 2-
1.8
HSO4 -
1.6
log Kb,m
1.4
1.2
1
0.8
HClO4 /H2 SO4
NaClO4 /Na2 SO4
0.6
NaClO4 media, Grenthe et al. ‘97
0.4
NaClO4 media, Lemire et al. ‘01
Calcu lation for a H2 SO4 mediu m
0.2
0
0
1
2
3
4
Im / mol kg-1
Figure 2.
173
100
%
155
253
123 139
149
263
167
191
218
316
271
132
334
289
299
351
369
387
405
236
449
467
485
0
120
140
160
180
200
220
240
260
280
300
320
340
360
380
400
420
440
460
480
m/z
500
32
Figure 3.
log β 1
4
3
2
log R
1
Sl
op
e1
0
-1
-2
-5
-4
-3
-2
log[SO4
-1
2-]
Figure 4.
7F
5D
0
1
Inorm R / a.u.
5D
0
2
570
7F
5D
0
0
580
590
600
610
620
630
640
650
7F
7F
5D
0
increasing [SO4 2-]
5D
0
560
7F
4
3
660
670
680
690
700
710
720
λemission / nm
33
Figure 5.
(a)
(b)
2.00-1.30 M Na +
4.5
100
2.00-1.30 M Na (Na2 SO4 /NaClO4)
Inorm R at 618 nm / a.u.
+
4
%
0.40-0.55 M Na (Na2 SO4 /NaClO4)
0.02-0.05 M H+ (H2SO4/HClO4)
+
3.5
3
Eu(S O 4)2-
0
Model A
100
Model B
2.5
Eu3+
EuS O 4+
50
%
0.40-0.55 M Na +
Eu3+
50
Eu(S O 4)2EuS O 4+
Model C
2
0
0.02-0.05 M H +
1.5
100
%
1
0.5
EuS O 4+
Eu3+
50
Eu(S O 4)2-
0
-7
-6
-5
-4
-3
-2
-1
0
-5
-4
-3
-2
-1
0
log[SO42-]
log[SO42-]
Figure 6.
4.5
Eu
4
Eu (TRLIL)
La
3.5
La (nanoESI-M S)
Eu in NaClO4 electro lyte
log β 1
3
Eu in Na2 SO4 electrolyte
2.5
2
1.5
1
0
1
2
3
Im / mol kg-1
34
Intensity / a.u.
Figure 7.
Eu(SO4 )2EuSO4+
Eu3+
560 570 580 590 600 610 620 630 640 650 660 670 680 690 700 710 720
λ emission / nm
35
Synopsis
Sulphate complexation of lanthanides is of great interest to predict the speciation of radionuclides in
natural environments. Thermodynamic constants were measured for the formation of the mono and
disulphate complexes of lanthanides(III) by TRLIL for Eu(III) and by nanoESI-MS for La(III). The
luminescence analysis of Eu(III) solutions suggested the formation of inner sphere complexes of
Ln(III) with monodentate SO42-.
H
0.5 M Na +
in Na 2SO4 /NaClO4
O
EuSO4+ (aq)
Ln
S
Eu(aq)3+
%
Eu(SO4)2- (aq)
-4
-3
-2
-1
log[SO 42-]
36